Update, April 1st 2015: http://www.zdnet.com/article/intel-claims-quantum-leap-in-fibre-optic-detectors/
Intel has developed silicon-based photoelectric detectors that could cut the cost of fiber-optic communications to a fraction of their current value.
That's nice I thought, and almost skipped on. But then I saw this part:
"Just to put it into context, a commercial [photodetector] that is used in telecoms typically sells for $200 to $300," Paniccia said. "We're talking devices that are probably an order or two in magnitude lower in cost." This prediction makes Intel's devices between 10 and 100 times cheaper than current photodetectors.
This means fiber optic to your house, which is already a paradigm changing event and already happening. But it gets better. These guys are talking about 40 to 200 gigabit per second speeds. For peanuts.
A DVD movie with all the trailers and special features is about 7 gigabytes, 56 gigabits. We're talking the downloading of a whole DVD movie in under two seconds.Your hard drive isn't nearly that fast, not even close.
This means a couple of things. First, bandwidth will soon be just amazingly cheaper. Fiber optic cable is -way- cheaper than coaxial cable or twisted pair copper, its just glass. The thing that's been holding back optical cable is the expense of these photodetectors. Intel can replace a $300 part with a $3.00 part, there's going to be optical cable getting laid down frickin' everywhere. That will mean your local cable company, local phone company etc. will be running optical to your house, and you'll probably be running optical -in- the house instead of CAT5 Ethernet or wireless. Sending really enormous amounts of data will be extremely cheap.
Second, cheap data means more data. The limiting step won't be transmission any more, it will be storage. You'll be able to have hi-def quality video in the security cameras at the bank machine. Count the pores on your nose quality. Live, real time, all the time. Youtube won't be shaky little squares on your screen anymore, it will be broadcast quality. On-line video games will be faster than lan-party, and the graphics will be Hollywood movie good.
Third, all this is going to crash together with the wireless boom going on right now and something will emerge that I can't even imagine. What could you do with every cell phone/Blackberry in the city having hi-def video and stereo sound recording and playback, plus GPS, plus as much on-board processing as a desk-top PC does today? Plus desk tops with 20 CPU cores, 200 gigs of RAM and multi-terabyte storage? Hooked to a 200 gb/sec fiber backbone?
This is going to be a Big Fricking Deal my friends, and it will be coming down the pipe to your house in under ten years. The internet, turbo-nitrous version.
Buy Cisco. They'll be selling some servers, I bet. Corning is going to sell some glass, too.
I understand you're a Calgary based lawyer who walks a thin line.
ReplyDeleteInteresting. LOL good thing I installed conduit throughout my home and shop.
ReplyDeleteBut now there seems to be a real lag in new processor development. I'm no expert, but seems to me that advances in cpu technology and architecture are slowing to a trickle in these recent years. Maybe we are approaching the limits of semiconductor integrated circuit technologies and will have to make a radical change in the technology soon?
---
anon said, "I understand you're a Calgary based lawyer who walks a thin line."
huh? wtf does this have to do with the post?
Hi Chris.
ReplyDeleteProcessor development seems to be focusing on parallelism these days. Good direction to go in, but its going to take a while for the software world to catch up. Meaning Mickeysoft of course, they being the elephant in the room.
They haven't even got up to speed on 64 bit architectures, which SGI was using back in the mid 1990s. I have an SGI Octane running IRIX on a 64 bit processor quite happily, brand new tech in maybe 1995.
I think a breakthrough like this fiber optic thing will jazz some life back into the tech industry. There are things you can do with a big fat pipe like this that can't be done at our current bandwidth level. We will see processors and especially storage technology speed -WAY- up to keep pace here.
Going to be a really interesting next 20 years, eh?